international journal of hydrogen energy 35 (2010) 4524–4533

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/he

Review

Development of large-scale hydrogen liquefaction processes

from 1898 to 2009

Songwut Krasae-in a,*, Jacob H. Stang b,1, Petter Neksa b,2

a
Norwegian University of Science and Technology, Kolbjorn Hejes vei 1d, NO-7491 Trondheim, Norway
b
SINTEF Energy Research AS, Kolbjorn Hejes vei 1d, NO-7465 Trondheim, Norway

article info abstract

Article history: This paper presents a review of the development of large-scale hydrogen liquefaction
Received 22 December 2009 processes throughout the world from 1898 to 2009. First, there is a concise literature review
Received in revised form including numerous past, present, and future designs given such as the first hydrogen
15 February 2010 liquefaction device, long time ago simple theoretical processes, today’s actual plants with
Accepted 21 February 2010 efficiencies 20–30%, a list of the capacity and location of every hydrogen liquefaction plant
Available online 26 March 2010 in the world, and some modern more efficient proposed conceptual plants with efficiencies
40–50%. After that, further information about the development and improvement potential
Keywords: of future large-scale liquid hydrogen liquefaction plants is given. It is found that every
Liquid hydrogen current plant is based on the pre-cooled Claude system, which is still the same as was 50
Hydrogen liquefier years ago with little improvement. Methods to resolve the challenges of the future plants
Large hydrogen liquefaction include proposing completely new configurations and efficient systems coupled with
Exergy efficiency improved efficiencies of the main system components such as compressors, expanders,
Comparison and heat exchangers. Finally, a summary and comparison of the process efficiencies are
described, including a newly proposed Multi-component Refrigerant (MR) system being
developed by NTNU and SINTEF Energy Research AS.
ª 2010 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved.

1. Introduction 50% [1–7]. This paper chronicles the development of systems

As hydrogen has shown promise as an important energy source
for use in future transportation vehicles, several hydrogen
research activities have been conducted since 1980 and espe-
cially since 2000. One of the challenges in creating a hydrogen
economy is the low efficiency of the current hydrogen lique-
faction plant cycles. Since 2000, there have been several papers
that have proposed conceptual plants with efficiencies up to 40–
from 1898 to 2009 and gives a comparison of several cycle effi-
ciencies for the future hydrogen plant developer. Hydrogen was
first liquefied in 1898 by a small device [8]. Some years later,
a pre-cooled Linde–Hampson system was used as the first
simple laboratory system to liquefy hydrogen. Around 1900,
more efficient laboratory systems were invented including the
Claude, pre-cooled Claude, and helium-refrigerated systems,
arranged in order of increasing efficiency [9]. Next, in 1957, the

* Corresponding author. Tel.: þ47 735 92991; fax: þ47 735 97214.
E-mail addresses: songwut.krasaein@ntnu.no, krasaein@hotmail.com (S. Krasae-in), jacob.stang@sintef.no (J.H. Stang),
petter.neksa@sintef.no (P. Neksa).
1
Tel.: þ47 735 98109.
2
Tel.: þ47 735 93923.
0360-3199/$ – see front matter ª 2010 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijhydene.2010.02.109
 international journal of hydrogen energy 35 (2010) 4524–4533
first few large hydrogen plants were built in the USA for the
growing petrochemical and aerospace industries and were
based on the pre-cooled Claude cycle with more complicated
systems that used liquid nitrogen as a pre-coolant to cool
hydrogen gas down to 193 C and hydrogen refrigeration
systems to further cool feed hydrogen gas to 253 C on a large
scale. Up to the present, almost all the large-scale plants in use
across the world today still employ nearly the same cycle as the
first few plants built in the US and have exergy efficiencies of
just 20–30%. This can be seen in the Ingolstadt plant installed in
Germany in 1991 [10]. Today, the most technologically
advanced plants available in the literature are located in Leuna,
Germany, and near Tokyo, Japan, which were commissioned in
2008; however, only a slight improvement of efficiency was
realized. Thus, there is potential for improvement.
2. Simple hydrogen liquefaction processes
Barron [9] illustrated the fundamental principles and how
these simple processes work very well.
2.1. The first hydrogen liquefaction system
In 1885, Michael Faraday published a paper regarding gas
liquefaction. At that time, his method was able to achieve
refrigeration temperatures down to 110 C using baths of
ether and solid carbon dioxide. Gases with boiling points
below that temperature, including hydrogen, were called
‘‘permanent gases’’ [11]. For the first time, the liquefaction of
hydrogen was achieved by Sir James Dewar in 1898 [8]. This
process utilized carbolic acid and liquid air for pre-cooling
compressed hydrogen at 180bars. The system was similar to
the one that Linde used for the liquefaction of air.
2.2. Theoretical liquefaction systems for hydrogen
In 1895, Carl von Linde and William Hampson invented
a simple liquefaction cycle to liquefy air. This cycle is called
the ‘Linde–Hampson cycle’. However, according to what was
explained by Barron [9], the systems that cannot be used to
liquefy hydrogen are the Linde–Hampson, Linde dual-
pressure, Cascade, and Heylandt systems. A liquid nitrogen,
pre-cooled Linde–Hampson system can be used to liquefy
hydrogen. This cycle is shown schematically by Barron [9].
2.3. Theoretical Claude system for hydrogen
In addition to liquefying air, the Claude cycle invented by
Georges Claude in 1902 can also be used to liquefy hydrogen
[9]. This cycle was a development some years after the first
Linde–Hampson cycle. There was an expansion engine in the
Claude cycle, which produced a temperature much lower than
the temperature generated by isenthalpic expansion as
proposed by Linde.
2.4. Theoretical pre-cooled Claude system for hydrogen
The performance is somewhat improved if a pre-cooling bath
of liquid nitrogen is used with the Claude system.
4525
Timmerhaus and Flynn [12] explained that if liquid nitrogen is
used for pre-cooling, one could achieve an exergy efficiency
50–70% higher than a pre-cooled Linde–Hampson cycle. Nandi
and Saragni [13] made a comparison of the two cycles and
found that the typical Figure of Merit (FOM) for the pre-cooled
Linde–Hampson cycle was lower than the standard pre-cooled
Claude. The Claude cycle, as explained by Nandi et al. [13], is
the basis for most other conventional liquefaction cycles. An
example of a modified pre-cooled Claude cycle in use today is
the hydrogen liquefaction plant in Ingolstadt near Munich,
Germany, as shown in Fig. 2, which has been in operation
since 1992 [10].
2.5. Helium-refrigerated hydrogen liquefaction system
A secondary helium-gas refrigerator can also be used to
liquefy hydrogen, as shown in Nandi and Sarangi [13] and
Barron [9], but this system has never been used in any actual
large-scale plants.
3. Current plants
Table 1 shows a list of all of the hydrogen liquefaction plants
in use around the world. In 1960, the first few liquid hydrogen
plants were built to support the Apollo program. In the
beginning of the 1960s there was a demand for US space
programs. The capacity installed up to 1965 was capable of
supplying the demand of NASA and others until 1977. In this
period, no additional plants were built, not least because of
the reduction of NASA’s space activities. Since 1977, this time
was mainly caused by the steadily increasing commercial
demand for liquid hydrogen. Today, there are more than 9
hydrogen liquefaction plants in the US with production rates
of 5–34 tons per day (TPD), 4 plants in Europe with capacities of
5–10 TPD, and 11 plants in Asia with capacities of 0.3–11.3 TPD.
Air Products supplies the largest quantity of liquid hydrogen
in North America, followed by Praxair. Today, liquid hydrogen
is used to reduce the cost of hydrogen distribution [14];
however, the current worldwide liquid hydrogen (LH2)
production capacity exceeds the market demand. Liquid
hydrogen demand and production today is the largest in North
America, which constitutes 84% of the world production. Of
the total production in the US, 33.5% is used in the petroleum
industry, 18.6% is for government aerospace, and the rest is
for other industries. Only 0.1% is used for fuel cells today [15].
3.1. Large-scale plants: Praxair, Air Products,
and Air Liquide
Praxair has five hydrogen liquefaction plants in the US today
with production rates between 6 and 35 TPD LH2. Typical
specific power consumptions are between 12.5 and 15 kW h/
kgLH2 [14]. Fig. 1 shows a Praxair LH2 process flow sheet. It
looks like the pre-cooled Claude cycle, but is more compli-
cated for the large-scale system. There are three heat
exchangers. The first heat exchanger is cooled by nitrogen gas
(GN2) and an external refrigeration system. The second heat
exchanger is cooled by liquid nitrogen (LN2) and some of the
H2 feed. The third is cooled by a hydrogen refrigeration system
4526 international journal of hydrogen energy 35 (2010) 4524–4533

Table 1 – Commercial hydrogen liquefaction plants worldwide.

Continent/Country Location Operated by Capacity (TPD) Commissioned in Still in operation

America
Canada Sarnia Air Products 30 1982 Yes
Canada Montreal Air Liquide 10 1986 Yes
Canada Inc.
Canada Becancour Air Liquide 12 1988 Yes
Canada Magog, Quebec BOC 15 1989 Yes
Canada Montreal BOC 14 1990 Yes
French Guyane Kourou Air Liquide 5 1990 Yes
USA Painsville Air Products 3a 1957 No
USA West Palm Beach Air Products 3.2a 1957 No
USA West Palm Beach Air Products 27a 1959 No
USA Mississippi Air Products 32.7a 1960 No
USA Ontario Praxair 20 1962 Yes
USA Sacramento Union Carbide, 54a 1964 No
Linde Div.
USA New Orleans Air Products 34 1977 Yes
USA New Orleans Air Products 34 1978 Yes
USA Niagara Falls Praxair 18 1981 Yes
USA Sacramento Air Products 6 1986 Yes
USA Niagara Falls Praxair 18 1989 Yes
USA Pace Air Products 30 1994 Yes
USA McIntosh Praxair 24 1995 Yes
USA East Chicago, IN Praxair 30 1997 Yes
Subtotal 300

Europe
France Lille Air Liquide 10 1987 Yes
Germany lngolstadt Linde 4.4 1991 Yes
Germany Leuna Linde 5 2008 Yes
Netherlands Rosenburg Air Products 5 1987 Yes
Subtotal 24.4

Asia
China Beijing CALT 0.6 1995 Yes
India Mahendragiri ISRO 0.3 1992 Yes
India India Asiatic Oxygen 1.2 – Yes
India Saggonda Andhra Sugars 1.2 2004 Yes
Japan Amagasaki Iwatani 1.2a 1978 No
Japan Tashiro MHI 0.6a 1984 No
Japan Akita Prefecture Tashiro 0.7 1985 Yes
Japan Oita Pacific Hydrogen 1.4 1986 Yes
Japan Tane-Ga-Shima Japan Liquid Hydrogen 1.4 1986 Yes
Japan Minamitane Japan Liquid Hydrogen 2.2 1987 Yes
Japan Kimitsu Air Products 0.3 2003 Yes
Japan Osaka Iwatani (Hydro Edge) 11.3 2006 Yes
Japan Tokyo Iwatani, built by Linde 10 2008 Yes
Subtotal 30.6

Worldwide 355

a Not included in the subtotal of the capacity for the year 2009.

that uses some of the feed to expand through turbines and the
Joule–Thomson (J–T) valve. The system is unique. Air Products
has four hydrogen liquefaction plants capable of producing
between 30 and 35 TPD LH2 in use in North America today. In
addition, they have two 5 TPD LH2 plants: one in Holland and
the other one in the USA. However, there is no literature about
Air Product’s technology. Air Liquide has a plant in France and
one in Canada, and both have capacities of about 10 TPD. Both
of these plants make use of the Claude cycle with hydrogen
used as the cycle fluid; however, there are no papers about Air
Liquide’s cycle. The best plant in the USA requires about
10 kW h/kgLH2 [14]. The LH2 production capacity is still greater
than the demand. It seems that every large-scale LH2 plant has
the cycle of LN2 as a pre-cooling process to cool hydrogen gas
from 25  C to 193  C and a hydrogen refrigeration system to
further cool hydrogen gas to 253  C.
3.2. Linde large-scale N2 pre-cooled Claude plant in
Ingolstadt
This plant used to be the largest German hydrogen liquefier.
The cycle is illustrated in Bracha et al. [10]. Feed hydrogen gas
obtained from an air separation plant is generated from
a steam reforming process using natural gas. Fig. 2 shows the
 international journal of hydrogen energy 35 (2010) 4524–4533
Fig. 1 – Praxair hydrogen liquefaction process (adapted
from [14]).
actual liquefier in the plant. The big, vertical tank nearby on
the left is the LN2 tank that the nitrogen liquefaction system
uses to liquefy nitrogen to pre-cool hydrogen inside the LH2
liquefier. All of the compressors are kept inside the machinery
building on the right. The leftmost tank is the LH2 storage tank
where liquefied hydrogen is kept for delivery. The tank is
vacuum insulated. Fig. 3(a) is for the other side. To minimize
the delivery cost, the hydrogen is delivered in liquid form by
truck. Fig. 3(b) demonstrates how LH2 is loaded from the
storage tank to the trailer.
3.3. The new Linde large-scale plant system in Leuna
Linde opened a second, 20 million Euro hydrogen liquefaction
plant in September 2007 in Leuna, as depicted in Figs. 5 and 6.
It is currently the newest and largest H2 liquefier plant in
Germany. The system with a new cycle as depicted in Fig. 4 is
similar to the existing plant in Ingolstadt depicted in Fig. 2, but
Fig. 2 – The location of Linde LH2 in Ingolstadt.
4527
is more efficient. There is an important difference in the
turbine arrangement between the plants in Leuna and Ingol-
stadt in that the plant in Leuna receives a single feed GH2
stream from an air separation plant. There is no recycled
hydrogen, and the ortho–para (o–p) conversions are put inside
heat exchangers.
4. Conceptual plants
From year 2000 to 2009, some researchers have proposed new
improved processes with exergy efficiencies between 40 and
50%. The details are given below.
4.1. Large-scale H2 liquefaction in combination with
liquefied natural gas (LNG) pre-cooling system
Kuendig et al. [1] conducted a study regarding the integration
of a pre-cooling LNG system to a new one like the Leuna N2
pre-cooled Claude system. The study concluded that using
LNG for pre-cooling in the hydrogen liquefaction process
would be extremely useful to decrease the power input and
the overall liquefier construction cost because the source
would be free. Compared to a conventional liquefaction
process, such as the one at Leuna using liquid nitrogen for pre-
cooling but with compression at ambient temperature, the
reduction would be from 10 to 4 kW h/kgLH2 [16]. However, this
process could only be used for hydrogen gas made from LNG,
and the plant would have to be located near a seaport.
4.2. Nitrogen pre-cooled Claude by Matsuda and
Nagami [2]
The World Energy NETwork (WE-NET) project [17] has suggested
building large-scale hydrogen liquefaction plants with lique-
faction capacities of 300 TPD. The plant is based on a Claude
cycle with nitrogen pre-cooling [2]. It appears that WE-NET’s
cycle is similar to the plant in Ingolstadt in that the nitrogen
cycle is used to pre-cool hydrogen from 25 to 193  C. Then, the
hydrogen cycle is used to cool from 193  C down to 253  C;
however, WE-NET’s cycle is more complicated and is specifi-
cally designed for greater capacity. There is a large N2 lique-
faction system to reliquify GN2 for the pre-cooling process.
4.3. Conceptual plant by Quack [3]
Quack [3] has made a conceptual design of a high-efficiency,
large-capacity liquefier for hydrogen. However, internal
process simulation tests run in a commercial software
package; SimSci/PROII by NTNU-SINTEF indicated that it was
not able to explicitly determine whether it has a high-effi-
ciency or not because the configuration of the proposed
propane refrigeration is impossible for low power consump-
tion. The software was checked for its reliability and accuracy
of process simulation. Also, the proposed helium–neon
refrigeration system consumes more power due to the fact that
helium–neon mixture has inferior refrigerant heat transfer
properties compared with hydrogen, which is commonly
found in use today in actual hydrogen liquefaction plants.
4528 international journal of hydrogen energy 35 (2010) 4524–4533

Fig. 3 – (a) Liquid hydrogen storage tank of Linde AG in Ingolstadt, (b) articulated train with semi-trailer equipped for liquid
hydrogen.

4.4. Conceptual plant with helium refrigeration cycle by
Kuzmenko et al. [4]
Before this, Beljakov et al. [18] successfully created a reliable,
high-efficiency, low-capacity hydrogen liquefier with a helium
refrigeration cycle. Later on, engineer Kuzmenko et al. [4] at
Open Joint-Stock Company used this concept to design
a liquefier. He made a conceptual study of building a medium-
capacity hydrogen liquefier with a helium refrigeration cycle;
however, it only produced a slight improvement from the
Ingolstadt plant’s efficiency.
4.5. MR refrigeration by Stang et al. [5]
A hydrogen liquefaction prototype laboratory unit was devel-
oped by NTNU–SINTEF. The process was based on using an MR
process for pre-cooling, as shown in the figure of Stang et al. [5].
The rig is under experiment by the author of this paper. With the
initial test, the hydrogen gas could be cooled by the MR refrig-
eration system from an ambient temperature of 25  C down to
near 193  C with the highest efficiency. Detailed experimental
results will be reported by the author in a future paper.
4.6. Helium refrigeration cycle by Shimko and Gardiner [6]
This is the design and construction of an estimated $2.6
million small-scale pilot plant (20 kg/h) that will be used for
hardware demonstration (will be finished in 2011) and as
a model for scaling to an estimated $39 million larger plant
(50 TPD) [6]. Simulations were performed using EXCEL and
REFPROP. Nevertheless, the efficiency is still lower than the
proposed NTNU–SINTEF system. Moreover, helium is not
suitable (hydrogen has better heat transfer properties) for
cooling GH2 from 193 to 253  C. If used, every component
such as compressors, expanders, and heat exchangers will
have to be bigger.

Fig. 4 – Process flow sheet of hydrogen liquefaction plant in Leuna [1].

 international journal of hydrogen energy 35 (2010) 4524–4533
Fig. 5 – Piston compressors of hydrogen liquefaction plant in Leuna (adapted from Finanzberichte.Linde.com, 2008).
4.7. Helium Joule–Brayton cascade system by Valenti [7]
Valenti [7] proposed an innovative, high-efficiency, large-scale
hydrogen liquefier that utilizes four cascaded helium Joule–
Brayton cycles. However, helium is not suitable for cooling
GH2 from 25 to 193  C and from 193 to 253  C due to its
inferior heat transfer properties compared to hydrogen.
Moreover, the cycle’s configuration itself to cool GH2 from
25  C to near 193  C makes it impossible to have low exergy
efficiency as reported. Also, internal simulation tests run in
PROII by NTNU-SINTEF indicated that the system is not
guaranteed to have a high-efficiency.
5. Development potential of large-scale LH2
plants
A potential efficiency increase in future hydrogen liquefaction
plants can be realized by the following means:
 Replacement of the J–T valve at the liquefaction stage by an
expansion turbine. An increase in the number and quality of
expansion turbines can minimize exergy losses.
 Reduction of the circulating mass flow or using a single H2
feed stream as used by the Leuna plant, Quack [3], and
Valenti [7]. By doing this, the last heat exchanger must be
designed to cool the hydrogen to the lowest possible
temperature, e.g. near 253  C, so there is no vapor fraction
after the expansion at the last J–T valve. A small ejector is
recommended to recover p-GH2 from the storage tank, the
same as the plant in Leuna.
 Operating with a refrigerant mixture for pre-cooling
hydrogen gas from 25 to 193  C. This way, pre-cooled
hydrogen gas and cold Multi-component Refrigerant (MR)
streams get closer. This new system is currently being
studied at NTNU–SINTEF [5], and the results will be pub-
lished soon.
 Another major factor influencing liquefier efficiency is the
feed gas input pressure. One alternative is to raise the
hydrogen output pressure of the preceding hydrogen
production plant, e.g. a high-pressure electrolysis process or
4529
a steam reforming plant. A good example is the 21 bar feed n-
H2 at the LH2 plants in Ingolstadt and Leuna. The higher the
feed pressure, the greater the liquefier’s efficiency. The
minimum liquefaction work is in conjunction with feed
pressure. The minimum feed pressure must not be below
15 bars because there could be hydrogen condensation
during the cooling process. If it is below 15 bars, more energy
is needed in liquefaction, and there will be more exergy loss.
 Most of the exergy losses in the hydrogen liquefaction
processes are dissipated through compressors. Therefore, it
is recommended for manufacturers to design new high-
efficiency compressors and expanders and design all
compressors in a way such that the suction temperatures
are reduced as done by Quack [3]. Also, it is recommended to
ventilate heat from the compressors as much as possible
during the compression process to reduce the exergy loss.
 Use aluminum plate-fin heat exchangers with maximum
effectiveness to reduce the exergy losses.
 If possible, construct plants near seaports for delivering LNG
to be used in the pre-cooling process. This will help signifi-
cantly reduce the plant size and energy consumption as
recommended by Kramer et al. [16] and Keundig et al. [1].
 A cost overview for the specific investment costs of
conventional liquefaction plants. When designing a large-
scale plant, the cost must be compared with other existing
plants. Inflation should be accounted for in current and old
plants. Companies who can offer cheap, large-scale
hydrogen liquefaction plants are Linde, Air Products, and
Praxair. Praxair has the largest hydrogen plant in the USA
with the lowest investment cost.
 Krewitt and Schmid [19] say that costs for liquefaction are
driven primarily by capital costs (today: 63%), then energy
costs (29%), and finally, O&M [14]. Energy costs on the other
hand are strong functions of the liquefier efficiency and
are less dependent on the production rate. In small plants,
energy and non-energy costs are comparable. In large-
scale plants, the energy costs become more important.
Krewitt and Schmid [19] also derived the following equa-
tion for the specific investment costs: Specific investment
cost for liquefier (V2000/kg/h) ¼ 828 313  (production
capacity, kg/h)0.48.
4530 international journal of hydrogen energy 35 (2010) 4524–4533
Fig. 6 – A Linde hydrogen (cold box) liquefier in Leuna (adapted from Linde-kryotechnik.ch, 2008).
 A method to decrease capital costs is to build plants on
a larger scale and use the effect of building multiple plants
of the same design. The following challenges for more
cost effective LH2 production systems are [14] system
modularization for traditional sized units, large-scale
equipment, higher efficiency compressors and expanders,
more efficient refrigeration, and lower cost high-efficiency
insulation.
The conclusions are the following:
 The problem with the current liquefaction systems is their
high-energy consumption. Every large-scale hydrogen
liquefaction plant is based on the pre-cooled Claude
system, which is still the same as 50 years ago with little
improvement. If it is possible to reduce from today’s energy
usage of 10 kW h/kgLH2 to around 5 kW h/kgLH2 which will
reduce electrical power consumption of the plant to be half
in the future, all of the compressors and motors in the
plant, which constitute the most expensive components,
could be reduced by 50%, which will also lead to cheaper
plants.
 Methods to resolve the challenges include proposing
completely new configurations and efficient systems
coupled with improved efficiencies of the main system
components such as compressors, expanders, and heat
exchangers.
 The development trend is that a lot of people have tried to
propose new better systems [1–7], but they are still nei-
ther more efficient nor realistic. Furthermore, compressor
and expander manufacturers must invent more efficient
machines.
6. Summary and comparison of hydrogen
liquefaction process efficiencies
Table 2 gives the summary and comparison. Feed hydrogen
flow is normal hydrogen at 1 atm, 25  C. FOM  100% ¼ (Ideal
liquefaction power/Actual system liquefaction power)  100%
or Exergy efficiency. The efficiencies of systems 3, 5, and 6 are
from Nandi and Sarangi [13]; the same systems have different
energy consumptions and exergy efficiencies because it
depends on the assumptions of the efficiencies of compres-
sors and expanders used in the systems. When making
a comparison between several different cycles and liquefiers,
Berstad et al.’s [20] comparison method is recommended. This
method, which is a direct comparison of liquefiers based on
the overall exergy efficiency and specific power consumption,
favors those with a higher portion of pre-compression. The
feed stream was assumed and calculated at 21 bars and 25  C
before going into any cycle/liquefier, which is identical to the
Ingolstadt plant. Every system is directly compared with the
Ingolstadt plant at a modified feed stream pressure of 21 bars.
The energy consumptions and exergy efficiencies of the
Ingolstadt, WE-NET, and Quack systems as shown in Table 2
were calculated by Berstad et al. [20].
The compression power reductions of the third, fourth,
fifth, and sixth hydrogen liquefaction systems in Table 2 are
0.9167, 0.9167, 0.2313, and 0.1026 kW h/kgLH2, respectively.
These are from the ideal H2 feed exergy reduction of 0.55, 0.55,
0.1388, and 0.0616 kW h/kgLH2, respectively. Make-up gas is
reversibly and isothermally (ideally) compressed from the feed
at 21 bars and 25  C to each cycle’s high side. This was all
calculated assuming a compression exergy efficiency of 60%.
 international journal of hydrogen energy 35 (2010) 4524–4533 4531

Table 2 – Summary and comparison of hydrogen liquefaction process efficiencies.

System with modified feed state: Original energy consumption Modified energy Modified exergy
normal hydrogen @21 bars, 25  C (kW h/kgLH2) consumption efficiency (%)
( kW h/kgLH2)

1. The thermodynamically ideal liquefaction system

Feed: 21 bars, 25  C, n-GH2
Output: 1 bar, 253  C, n-LH2 – 2.178 100
Output: 1 bars, 253  C, 99.8% p-LH2 – 2.890 100

2. Theoretical simple Linde–Hampson system [8]. – – –

*Cannot liquefy hydrogen
3. Theoretical pre-cooled Linde–Hampson [8,13]. 64.5–71.7 63.6–70.8 3.0–3.4
Output: 1 bar, 253  C, n-LH2
4. Theoretical Claude system [8,13]. Less than the pre-cooled Claude
Output: 1 bar, 253  C, n-LH2
5. Theoretical pre-cooled Claude system [8,13]. 24.8–35.0 24.6–34.8 6.2–8.8
Output: 1 bar, 253  C, n-LH2
6. Theoretical 29.3–49.5 29.2–49.4 4.4–7.4
helium-refrigerated system [8,13].
Output: 1 bar, 253  C, n-LH2
7. Large-scale Praxair plant system [14]. z12–15 19–24
Output: z1 bar, 253  C, 95% p-LH2
8. Large-scale Air Products plant system [14].
Output: z1 bar, 253  C, 95% p-LH2
9. Large-scale Air Linde plant system [14].
Output: z1 bar, 253  C, 95% p-LH2
10. Large-scale plant, Claude system in Ingolstadt 13.58 21.0
on stream in 1994 by Bracha et al. [10]. Output:
1.3 bars, 253  C, 95% p-LH2
11. WE-NET: Nitrogen pre-cooled large-scale Claude
plant by Matsuda and Nagami [2] Output: 1.3
bars, 253  C, 95% p-LH2
1) Hydrogen Claude z8.5 N/A N/A
2) Helium Brayton
3) Basic neon
4) Neon with cold pump 7.0 41.3

12. Large-scale conceptual plant by Quack [3]

Output: 1 bar, 253  C, 99.8% p-LH2
1) Without pressure drop in calculation 7.0 5.49 52.6
2) With pressure drop in calculation z7.3 N/A N/A

13. Four helium Joule–Brayton cascade cycle by 5.04 5.76 50.2

Valenti [7]. Output: 1.5 bars, 253  C, 99.8% p-LH2

For cycles 7–9, the hydrogen feed pressure was 21 bars, the
same as Ingolstadt’s. Thus the energy consumption was the
same. With Valenti’s system, GH2 compression must be made
from 21 bar supply feed to 60 bars; therefore, there is an
increased consumption of 0.72 kW h/kgLH2 with an assumed
60% exergy efficiency from the ideal H2 feed exergy increase of
0.43 kW h/kgLH2. Finally, all of the system exergy efficiencies
were calculated by comparing with an ideal energy consump-
tion of 2.89 kW h/kgLH2; however, systems 3–6 were calculated
using an ideal energy consumption of 2.178 kW h/kgLH2.
Fig. 7 contains the information shown in Table 2. From the
data, the theoretical pre-cooled Linde–Hampson system was
the first imaginary system invented a long time ago, and its
exergy efficiency is the lowest. After that, the second was the
theoretical helium-refrigerated system, which is followed by
the theoretical pre-cooled Claude system. All have a very low
yield: e.g. 10% after expansion. The theoretical systems
mentioned have never been used to liquefy hydrogen on
a large-scale production. They were just small-scale labora-
tory systems. Next, Ingolstadt and Praxair brought this
concept to invent real plants. Today, actual large-scale
hydrogen liquefaction plants, e.g. Praxair, Air Products, and
Air Liquids plants in the USA, energy consumptions are
reported to be between 12 and 15 kW h/kgLH2 [14]. Baker and
Shaner’s [21] was the first conceptual plant, and it had
the lowest efficiency. The conceptual large-scale systems
proposed by Matsuda and Nagami [2], Quack [3], and Valenti
[7] were designed later. Recently, the efficiency of the Leuna
plant (with energy consumption less than 13.58 kW h/kgLH2) is
a little better than Ingolstadt is assumed here. Quack’s process
reports the best cycle exergy efficiency at 5.76 kW h/kgLH2. The
best plant in the USA today is reported to require 10 kW h/
kgLH2 [14], but it is not known where. A simulated 50 TPD large-
scale Shimko plant, which is a helium refrigeration system
with a hydrogen feed at 21 bars, is reported at 8.7 kW h/kgLH2.
The proposed large-scale MR system is 5.35 kW h/kgLH2 as
depicted. The ideal theoretical minimum value is 2.89 kW h/
kgLH2. For the process with LNG for pre-cooling studied by
Kuendig et al. [1], the power consumption is reported by
Kramer et al. [16] to be 4 kW h/kgLH2. Thus, the overall
4532 international journal of hydrogen energy 35 (2010) 4524–4533

80
Theoretical cycle long time ago

Electrical energy consumption (kWh/kgLH2)

70 Theoretical pre-cooled Linde-Hampson system
Today real large plant
Conceptual large plant
60 The proposed
large-scale MR system
50 Theoretical helium-refrigerated system

40
Theoretical pre-cooled Claude system
30
Praxair, Air Products, and Air Liquids from 1957 to 2009
Linde plant in Leuna in 2007
20 The best US plant up to 2009
Baker (1978)
Linde plant in The proposed large-scale MR system (2010)
10
Ingolstadt in 1992
Shimko (2007)
Matsuda (1998) Valenti (2008) Quack (2002)
0
0 10 20 30 40 50 60
Overall cycle exergy efficiency (%)

Fig. 7 – Comparison of hydrogen liquefaction process efficiencies by assuming that all processes are with uniform feed
pressure equal to that of Ingolstadt plant at 21 bars.

efficiency, compared with the ideal process, is [(2.89 kW h/ [2] Matsuda H, Nagami M. Study of large hydrogen liquefaction
kgLH2)/(4 kW h/kgLH2)]  100 ¼ 72%, which is the highest with process. Kanagawa, Japan: Nippon Sanso Corp. WE-NET:
respect to all current systems. However, this is not shown in summary of annual reports. Available from: http://www.
enaa.or.jp/WE-NET/ronbun/1997/e5/sanso1997.html; 1998
Fig. 7 because the process is cooled by free LNG, not by the
[accessed 05.06.09].
system itself. Completely new approaches for low tempera- [3] Quack H. Conceptual design of a high efficiency large
ture refrigeration are magnetic refrigerators and acoustic capacity hydrogen liquefier. Adv Cry Eng 2002;47:255–63.
refrigerators. Magneto caloric cooling may reduce liquefaction [4] Kuzmenko IF, Morkovkin IM, Gurov EI. Concept of building
energy to 5.0 kW h/kgLH2 [22]; however, this may only be for medium-capacity hydrogen liquefiers with helium
small-scale to medium-scale plants. All of the literature refrigeration cycle. Chem Pet Eng 2004;40(Nos: 1–2).
related to magnetic cooling has been reorted on small-scale [5] Stang J, Neksa P, Brendeng E. On the design of an efficient
hydrogen liquefaction process. WHEC 2006; 16/13–16 June
hydrogen plants. Nobody thinks such a system is realistic in
2006 – Lyon France.
large-scale systems. [6] Shimko M, Gardiner M. Innovative hydrogen liquefaction
cycle. Available from: http://www.hydrogen.energy.gov/
pdfs/progress08/iii_7_shimko.pdf; 2007 [accessed 14.06.09].
7. Conclusion [7] Valenti G. Proposal of an innovative, high-efficiency, large-
scale hydrogen liquefier. Int J Hydrogen Energy 2008;33(12):
Today large hydrogen liquefaction plants have exergy effi- 3116–21.
ciencies of just 20–30%; thus, there is potential for improve- [8] Dewar J. Liquid hydrogen. Science 1898;8(183):3–6.
[9] Barron RF. Cryogenic systems. Oxford, United Kingdom:
ment. From 1998 to 2008, some conceptual plants have been
Oxford University Press; 1966.
proposed with reported efficiencies of 40–50%. Finally, in the [10] Bracha M, Lorenz GA, Wanner M. Large-scale hydrogen
year 2010, NTNU and SINTEF Energy Research AS will propose liquefaction in Germany. Int J Hydrogen Energy 1994;19(1):53–9.
a new MR system with an efficiency greater than 50%. Details [11] Foerg W. History of cryogenics: the epoch of the pioneers
of the proposed system will be reported in upcoming papers. from the beginning to the year 1911. Int J Refrig 2002;25(3):
283–92.
[12] Timmerhaus KD, Flynn TM. Cryogenic process engineering.
New York: Springer; 1989.
Acknowledgements [13] Nandi TK, Sarangi S. Performance and optimization of
hydrogen liquefaction cycles. Int J Hydrogen Energy 1993;
The author wishes to thank the Department of Energy and 18(2):131–9.
[14] Drnevich R. Hydrogen delivery – liquefaction & compression.
Process Engineering, Norwegian University of Science and
Praxair, strategic initiatives for hydrogen delivery workshop
Technology for a research fellow grant.
– May 7, 2003. Available from: http://www1.eere.energy.gov/
hydrogenandfuelcells/pdfs/liquefaction_comp_pres_praxair.
pdf [accessed 20.5.09].
references [15] Franser D. Solutions for hydrogen solutions for hydrogen
storage and distribution by dynetek industries Ltd. The PEI
Wind-Hydrogen Symposium June 22–24, 2003. Available
[1] Kuendig A, Loehlein K, Kramer GJ, Huijsmans J. Large scale from: http://www.gov.pe.ca/photos/original/dev_solutions.
hydrogen liquefaction in combination with LNG re- pdf [accessed 04.03.08].
gasification. Available from: http://www.cder.dz/A2H2/ [16] Kramer GJ, Huijsmans J, Austgen D. Clean and green
Medias/Download; 2006 [accessed 05.06.09]. hydrogen. 16th World hydrogen energy conference, 16/13–16
 international journal of hydrogen energy 35 (2010) 4524–4533
June 2006 – Lyon, France. Available from: http://www.shell.
com/static/hydrogen-en/downloads/speeches/speech_
clean_green_hydrogen_1.pdf [accessed 05.06.09].
[17] Mitsugi C, Harumi A, Kenzo F. WE-NET: Japanese hydrogen
program. Int J Hydrogen Energy 1998;23(3):159–65.
[18] Beljakov V, Krakovsky BD, Popov OM. Low capacity hydrogen
liquefier with helium cycle. In: Proc Cryogenics 2000; Praha:
p. 158–61.
[19] Krewitt W, Schmid S. Common information database – fuel
cell technologies and hydrogen production/distribution
4533
options. Available from: http://www.dlr.de/fk/Portaldata/40/
Resources/dokumente/publikationen/2005-09-02_CASCADE_
D1.1_fin.pdf; 2005 [accessed 14.06.09].
[20] Berstad D, Stang J, Neksa P. Comparison criteria for large-
scale hydrogen liquefaction processes. Int J Hydrogen Energy
2009;34(3):1560–8.
[21] Baker CR, Shaner RL. A study of the efficiency of hydrogen
liquefaction process. Int J Hydrogen Energy 1978;3(3):
321–34.
[22] Dutton G. Hydrogen transitions – a UK perspective; 2003.